linux-kernel - Re: [PATCH 18/18] mfd: wm8400-core: Make it explicitly non-modular

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 28 Dec 2018 09:42:36 +0000
From:   Charles Keepax <ckeepax@...nsource.cirrus.com>
To:     Paul Gortmaker <paul.gortmaker@...driver.com>
CC:     Lee Jones <lee.jones@...aro.org>, <linux-kernel@...r.kernel.org>,
        Mark Brown <broonie@...nsource.wolfsonmicro.com>,
        <patches@...nsource.cirrus.com>,
        Linus Walleij <linus.walleij@...aro.org>
Subject: Re: [PATCH 18/18] mfd: wm8400-core: Make it explicitly non-modular

On Fri, Dec 21, 2018 at 10:55:16AM -0500, Paul Gortmaker wrote:
> [Re: [PATCH 18/18] mfd: wm8400-core: Make it explicitly non-modular] On 19/12/2018 (Wed 09:17) Charles Keepax wrote:
> 
> > On Mon, Dec 17, 2018 at 03:31:28PM -0500, Paul Gortmaker wrote:
> > > -MODULE_DEVICE_TABLE(i2c, wm8400_i2c_id);
> > >  
> > >  static struct i2c_driver wm8400_i2c_driver = {
> > >  	.driver = {
> > > @@ -161,7 +160,7 @@ static struct i2c_driver wm8400_i2c_driver = {
> > >  };
> > >  #endif
> > 
> > Do we not want to add suppress_bind_attrs into the i2c_driver
> > struct here?
> 
> We can add one if you/maintainers want one, but if you look at the
> original patch, this driver was using the more classic/legacy case of
> subsys_init() vs. platform_driver_register() used in other drivers.
> 
> Not adding a suppress_bind_attrs here was intentional, since I'd decided
> to put in the unbind entries for code that used platform_driver_register()
> where the author had created the .remove code, on the assumption that they
> had put some thought into the process of unbind/remove - to make it
> explicit that unbind is now disabled.
> 
> To be clear, using the subsys_init() doesn't implicitly disable unbind.
> However, there are lots of non-modular drivers out there; ones I've not
> even touched, and to start a project to add an unbind disable to them
> all is beyond the scope of the goals I've listed in the 00/18 preamble.
> 
> I'd hope maybe we can revisit the global default setting for non-modular
> code someday - to make non-modules opt-in instead of opt-out, and
> achieve better consistency from one driver to the next, without having
> to add a new .driver sub-struct to each file for the suppress entry.
> 
> I think LinusW hinted at in an earlier email in this ongoing review,
> that the default setting didn't quite make sense to him either.  But in
> any case, that is a separate discussion for another time and place.
> 
> Let me know if you explicitly want one added, otherwise I'll just leave
> the .remove + .suppress_bind_attrs pairing as described above.
> 

Nah its ok, if you are specifically not doing this one I think I
am ok with it. Just seemed out of place compared to the the
others:

Acked-by: Charles Keepax <ckeepax@...nsource.cirrus.com>

Thanks,
Charles